Maintaining high standards for user safety during daily railway operations is crucial for railway managers. To aid in this endeavor, top or side-view cameras and GPS positioning systems have facilitated progress toward automating periodic inspections of defective features and assessing the deteriorating status of railway components. However, collecting data on deteriorated status can be time-consuming and requires repeated data acquisition because of the extreme temporal occurrence imbalance. In supervised learning, thousands of paired data sets containing defective raw images and annotated labels are required. Concretely, the one-class classification approach offers the advantage of requiring quite a few anomalous images to optimize parameters for training large normal images. The deeper fully-convolutional data descriptions (FCDDs) were applicable to several damage data sets of concrete/steel components in structures, and fallen tree, and wooden building collapse in disasters. However, it is not yet known to feasible to railway components. In this study, we devised a prognostic discriminator pipeline to automate one class classification using the augmented deeper FCDDs for defective railway components. We also performed sensitivity analysis of the mixture and erasing augmentations, and the deeper backbone rather than the shallow baseline of convolutional neural network (CNN) with 27 layers. Furthermore, we visualized defective railway features by using transposed Gaussian upsampling. We demonstrated our application to railway inspection using a video acquisition dataset that contains wooden sleeper deterioration. Finally, we examined the usability of our approach for prognostic monitoring and fu ture work on railway component inspection.
Rural Railway Prognostics, Automated Visual Inspection, Decayed Wooden sleeper, One-class Classification, Damage Explanation
Chandran, P., Asber, J., Thiery, F., Odelius, J., & Rantatalo, M. (2021). An investigation of railway fastener detec tion using image processing and augmented deep learn ing. Sustainability, 13(21).
C. Shorten, T. K. (2019). A survey on image data augmenta tion for deep learning. In Proceedings of the 35th inter national conference on machine learning (Vol. 6:60).
Devries, T., & Taylor, G. W. (2017). Improved regularization of convolutional neural networks with cutout. ArXiv, abs/1708.04552.
Evans, A. W. (2011). Fatal train accidents on Europe’s rail ways: 1980–2009. Accident Analysis and Prevention, 43(1), 391-401.
Evans, A. W. (2020). Fatal train accidents on Europe’s rail ways: 1980–2019 (Tech. Rep.). Centre for Transport Studies, Imperial College London.
Hsieh, C.-C., Hsu, T.-Y., & Huang, W.-H. (2022). An online rail track fastener classification system based on YOLO models. Sensors, 22(24).
Hsieh, C.-C., Lin, Y.-W., Tsai, L.-H., Huang, W.-H., Hsieh, S.-L., & Hung, W.-H. (2020). Offline deep-learning based defective track fastener detection and inspection system. Sensors and Materials, 32(10), 3429.
Ji, A., Woo, W. L., Wong, E. W. L., & Quek, Y. T. (2021). Rail track condition monitoring: A review on deep learning approaches. Intelligence and Robotics, 1(2), 151–175.
Liznerski, P., Ruff, L., Vandermeulen, R. A., Franks, B. J., Kloft, M., & Muller, K.-R. (2021). Explainable deep ̈ one-class classification. In The international confer ence on learning representations(ICLR).
Mi, Z., Chen, R., & Zhao, S. (2023). Research on steel rail surface defects detection based on improved YOLOv4 network. Frontiers in Neurorobotics, 17.
Miwa, M. (2019). Railway maintenance transformed by Nu merical Optimizer. In Mathematical systems user con ference 2019.
Oyama, T., & Miwa, M. (2022). Applying probabilistic math ematical modeling approach and ai technique to inves tigate serious train accidents in japan. Sustainability Analytics and Modeling, 2, 2667-2596.
Ribeiro, M. T., Singh, S., & Guestrin, C. (2016). ”why should i trust you?”: Explaining the predictions of any classifier. In Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining (p. 1135–1144). Association for Comput ing Machinery.
Ruff, L., Vandermeulen, R. A., Franks, B. J., Muller, K.-R., ̈ & Kloft, M. (2021). Rethinking assumptions in deep anomaly detection. In The international conference on machine learning (ICML).
Selvaraju, R. R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., & Batra, D. (2017). Grad-CAM: Visual ex planations from deep networks via gradient-based lo calization. In 2017 IEEE international conference on computer vision (ICCV) (p. 618-626).
Takahashi, R., Matsubara, T., & Uehara, K. (2018). Ri cap: Random image cropping and patching data aug mentation for deep cnns. In Proceedings of the 10th asian conference on machine learning (Vol. 95, pp. 786–798). PMLR.
Tang, R., De Donato, L., Besinovic, N., Flammini, F., Goverde, R. M., Lin, Z., . . . Wang, Z. (2022). A literature review of artificial intelligence applications in railway systems. Transportation Research Part C: Emerging Technologies, 140, 103679.
Tango, K., Ohkawa, T., Furuta, R., & Sato, Y. (2022). Back ground mixup data augmentation for hand and object in-contact detection.
Yasuno, T., Okano, M., & Fujii, J. (2023). One-class dam age detector using deeper fully convolutional data de scriptions for civil application. Advances in Artificial Intelligence and Machine Learning, 3(2), 996-1011.
Zeiler, M. D., & Fergus, R. (2013). Visualizing and under standing convolutional networks.
Zhang, H., Cisse, M., Dauphin, Y. N., & Lopez-Paz, D. (2018). mixup: Beyond empirical risk minimization. In International conference on learning representations.
Zhong, Z., Zheng, L., Kang, G., Li, S., & Yang, Y. (2020). Random erasing data augmentation. In The 34th aaai conference on artificial intelligence (AAAI-20).
Zhou, B., Khosla, A., Lapedriza, A., Oliva, A., & Torralba, A. (2015). Learning deep features for discriminative localization.
This work is licensed under a Creative Commons Attribution 3.0 Unported License.